Data Visualisation: Telling Visual Stories From Data
Bevan Koopman
11 /04/2017
Anscombe’s quartet
anscombe_m <- read.csv("anscombe_m.csv")
summary(anscombe_m)## Dataset x y
## A:11 Min. : 4 Min. : 3.100
## B:11 1st Qu.: 7 1st Qu.: 6.117
## C:11 Median : 8 Median : 7.520
## D:11 Mean : 9 Mean : 7.501
## 3rd Qu.:11 3rd Qu.: 8.748
## Max. :19 Max. :12.740Checks the mean, variance for each dataset
aggregate(y ~ Dataset, anscombe_m, mean)## Dataset y
## 1 A 7.500909
## 2 B 7.500909
## 3 C 7.500000
## 4 D 7.500909aggregate(x ~ Dataset, anscombe_m, mean)## Dataset x
## 1 A 9
## 2 B 9
## 3 C 9
## 4 D 9aggregate(y ~ Dataset, anscombe_m, var)## Dataset y
## 1 A 4.127269
## 2 B 4.127629
## 3 C 4.122620
## 4 D 4.123249aggregate(x ~ Dataset, anscombe_m, var)## Dataset x
## 1 A 11
## 2 B 11
## 3 C 11
## 4 D 11for(d in c("A", "B", "C", "D")) {
print(cor(anscombe_m[anscombe_m$Dataset==d,]$x, anscombe_m[anscombe_m$Dataset==d,]$y))
}## [1] 0.8164205
## [1] 0.8162365
## [1] 0.8162867
## [1] 0.8165214Now plot the four different data sets
ggplot(anscombe_m, aes(x=x, y=y)) + geom_point(size=3) + facet_wrap(~Dataset) 
Grammar of graphical elements
1. The Data layer
Have a look at the diamonds dataset.
summary(diamonds)## carat cut color clarity
## Min. :0.2000 Fair : 1610 D: 6775 SI1 :13065
## 1st Qu.:0.4000 Good : 4906 E: 9797 VS2 :12258
## Median :0.7000 Very Good:12082 F: 9542 SI2 : 9194
## Mean :0.7979 Premium :13791 G:11292 VS1 : 8171
## 3rd Qu.:1.0400 Ideal :21551 H: 8304 VVS2 : 5066
## Max. :5.0100 I: 5422 VVS1 : 3655
## J: 2808 (Other): 2531
## depth table price x
## Min. :43.00 Min. :43.00 Min. : 326 Min. : 0.000
## 1st Qu.:61.00 1st Qu.:56.00 1st Qu.: 950 1st Qu.: 4.710
## Median :61.80 Median :57.00 Median : 2401 Median : 5.700
## Mean :61.75 Mean :57.46 Mean : 3933 Mean : 5.731
## 3rd Qu.:62.50 3rd Qu.:59.00 3rd Qu.: 5324 3rd Qu.: 6.540
## Max. :79.00 Max. :95.00 Max. :18823 Max. :10.740
##
## y z
## Min. : 0.000 Min. : 0.000
## 1st Qu.: 4.720 1st Qu.: 2.910
## Median : 5.710 Median : 3.530
## Mean : 5.735 Mean : 3.539
## 3rd Qu.: 6.540 3rd Qu.: 4.040
## Max. :58.900 Max. :31.800
## 2. The Aesthetics layer
Now we map parts of the data to some graphical elements (x, y, color, size, etc)
ggplot(diamonds, aes(x=carat, y=price)) 
Note, that there is no notion at the moment of the types of plot - that is the next layer!
3. The Geometries layer
Now we define a geometry (this case, a scatter geometry):
ggplot(diamonds, aes(x=carat, y=price)) + geom_point(alpha=.4)
4. The Facet layer
Now lets split the plot up by clarity:
ggplot(diamonds, aes(x=carat, y=price)) + geom_jitter(alpha=.4) + facet_wrap(~clarity)
5. The Statistics layer
Now let’s add an graphical element that aids in understanding, namely a linear model:
ggplot(diamonds, aes(x=carat, y=price)) + geom_jitter(alpha=.4) + facet_wrap(~clarity) + geom_smooth(method="lm", se=F)
6. The Coordinates layer
Now let’s ajust the scales and axis levels. Specifically, we’ll restrict x-axis carat to be between 0 and 3, report x-axis carat in 0.5 increments, change y-axis price to be in dollar format and provide x and y axis labels.
library(scales) # used to print things in dollar ($) format
ggplot(diamonds, aes(x=carat, y=price)) + geom_jitter(alpha=.4) + facet_wrap(~clarity) + geom_smooth(method="lm", se=F) + scale_x_continuous(limits=c(0,3), breaks = seq(0,3, 0.5)) + scale_y_continuous(labels = dollar) + xlab("Mass (carat)") + ylab("Price (US$)")## Warning: Removed 32 rows containing non-finite values (stat_smooth).## Warning: Removed 35 rows containing missing values (geom_point).
6. The Theme layer
Now just change the look and feel:
ggplot(diamonds, aes(x=carat, y=price)) + geom_jitter(alpha=.4) + facet_wrap(~clarity) + geom_smooth(method="lm", se=F) + scale_x_continuous(limits=c(0,3), breaks = seq(0,3, 0.5)) + scale_y_continuous(labels = dollar) + xlab("Mass (carat)") + ylab("Price (US$)") + theme_minimal()## Warning: Removed 32 rows containing non-finite values (stat_smooth).## Warning: Removed 35 rows containing missing values (geom_point).
Data that supports Visualisation
Have a look at the airquality dataset.
summary(airquality)## Ozone Solar.R Wind Temp
## Min. : 1.00 Min. : 7.0 Min. : 1.700 Min. :56.00
## 1st Qu.: 18.00 1st Qu.:115.8 1st Qu.: 7.400 1st Qu.:72.00
## Median : 31.50 Median :205.0 Median : 9.700 Median :79.00
## Mean : 42.13 Mean :185.9 Mean : 9.958 Mean :77.88
## 3rd Qu.: 63.25 3rd Qu.:258.8 3rd Qu.:11.500 3rd Qu.:85.00
## Max. :168.00 Max. :334.0 Max. :20.700 Max. :97.00
## NA's :37 NA's :7
## Month Day
## Min. :5.000 Min. : 1.0
## 1st Qu.:6.000 1st Qu.: 8.0
## Median :7.000 Median :16.0
## Mean :6.993 Mean :15.8
## 3rd Qu.:8.000 3rd Qu.:23.0
## Max. :9.000 Max. :31.0
## Get the reshape2 package (not reshape).
require(reshape2)Look at the top 10 rows of the data:
head(airquality)## Ozone Solar.R Wind Temp Month Day
## 1 41 190 7.4 67 5 1
## 2 36 118 8.0 72 5 2
## 3 12 149 12.6 74 5 3
## 4 18 313 11.5 62 5 4
## 5 NA NA 14.3 56 5 5
## 6 28 NA 14.9 66 5 6Transform to long format.
head(melt(airquality))## No id variables; using all as measure variablesTranform to long format but keep Month and Day as columns.
long_airquality <- melt(airquality, id.vars=c('Month', 'Day'), variable.name = "Measurement")
head(long_airquality)## Month Day Measurement value
## 1 5 1 Ozone 41
## 2 5 2 Ozone 36
## 3 5 3 Ozone 12
## 4 5 4 Ozone 18
## 5 5 5 Ozone NA
## 6 5 6 Ozone 28Now the data is in a easier format for plotting.
ggplot(long_airquality, aes(x=Month, y=value)) + geom_point() + facet_wrap(~Measurement, scales="free")## Warning: Removed 44 rows containing missing values (geom_point).
Calculate a montly average
agg_long_airquality <- aggregate(value ~ Month + Measurement, long_airquality, mean)
ggplot(agg_long_airquality, aes(x=Month, y=value)) + geom_point() + geom_line() + facet_wrap(~Measurement, scales="free")
But rather than calculate the average, why not just use a boxplot:
ggplot(long_airquality, aes(x=Month, y=value)) + geom_jitter() + facet_wrap(~Measurement, scales="free") # the old way## Warning: Removed 44 rows containing missing values (geom_point).
ggplot(long_airquality, aes(x=Month, y=value)) + geom_jitter() + facet_wrap(~Measurement, scales="free") + stat_summary(fun.y="mean", color="red", size=3, geom="point")## Warning: Removed 44 rows containing non-finite values (stat_summary).
## Warning: Removed 44 rows containing missing values (geom_point).
ggplot(long_airquality, aes(x=factor(Month), y=value)) + geom_boxplot(varwidth = T) + facet_wrap(~Measurement, scales="free") + scale_x_discrete()## Warning: Removed 44 rows containing non-finite values (stat_boxplot).
Different Plot Types
Let’s return to diamonds dataset and looks through a bunch of different plot types and what they show.
Just a reminder of the diamonds dataset:
summary(diamonds)## carat cut color clarity
## Min. :0.2000 Fair : 1610 D: 6775 SI1 :13065
## 1st Qu.:0.4000 Good : 4906 E: 9797 VS2 :12258
## Median :0.7000 Very Good:12082 F: 9542 SI2 : 9194
## Mean :0.7979 Premium :13791 G:11292 VS1 : 8171
## 3rd Qu.:1.0400 Ideal :21551 H: 8304 VVS2 : 5066
## Max. :5.0100 I: 5422 VVS1 : 3655
## J: 2808 (Other): 2531
## depth table price x
## Min. :43.00 Min. :43.00 Min. : 326 Min. : 0.000
## 1st Qu.:61.00 1st Qu.:56.00 1st Qu.: 950 1st Qu.: 4.710
## Median :61.80 Median :57.00 Median : 2401 Median : 5.700
## Mean :61.75 Mean :57.46 Mean : 3933 Mean : 5.731
## 3rd Qu.:62.50 3rd Qu.:59.00 3rd Qu.: 5324 3rd Qu.: 6.540
## Max. :79.00 Max. :95.00 Max. :18823 Max. :10.740
##
## y z
## Min. : 0.000 Min. : 0.000
## 1st Qu.: 4.720 1st Qu.: 2.910
## Median : 5.710 Median : 3.530
## Mean : 5.735 Mean : 3.539
## 3rd Qu.: 6.540 3rd Qu.: 4.040
## Max. :58.900 Max. :31.800
## The scatter plot
ggplot(diamonds, aes(x=carat, y=price)) + geom_point()
Now let’s add a visual distinction (as color) for diamond clarity:
ggplot(diamonds, aes(x=carat, y=price, color=clarity)) + geom_point()
Now lets add another categorical dimension for the quality of the cut as the size:
ggplot(diamonds, aes(x=carat, y=price, color=clarity, size=cut)) + geom_point()## Warning: Using size for a discrete variable is not advised.
Trends
Now let’s apply a smoothing trend:
ggplot(diamonds, aes(x=carat, y=price, color=clarity)) + geom_point() + geom_smooth(se=F)## `geom_smooth()` using method = 'gam'
ggplot(diamonds, aes(x=carat, y=price, color=clarity)) + geom_smooth(se=F)## `geom_smooth()` using method = 'gam'
Faceting
The plot is pretty busy so lets rather divide things up according to the categorical data.
First divide by cut:
ggplot(diamonds, aes(x=carat, y=price, color=clarity)) + geom_point() + facet_wrap(~cut)
Maybe by clarity instead:
ggplot(diamonds, aes(x=carat, y=price, color=cut)) + geom_point() + facet_wrap(~clarity)
And add a further facet for color:
ggplot(diamonds, aes(x=carat, y=price, color=cut)) + geom_point() + facet_grid(color~clarity) + scale_color_grey()
Barplot
Barplot are used for make comparisons. Let’s have a look at some climate data.
global_temps <- read.table("647_Global_Temperature_Data_File.txt", col.names = c("Year", "TempChange", "TempChange5Year"))The data shows change in global surface temperature for a 136-year period, relative to 1951-1980 average temperatures.
ggplot(global_temps, aes(x=Year, y=TempChange)) + geom_bar(stat="identity")
Let’s add the bounds for the comparitive years:
ggplot(global_temps, aes(x=Year, y=TempChange)) + geom_bar(stat="identity") + geom_vline(xintercept=c(1951,1980), color="red")
Line plots
Line plots are designed to show trends.
Let’s look at the change in temps over the years:
ggplot(global_temps, aes(x=Year, y=TempChange)) + geom_line()
Sometimes we want a trend and to compare indvidual years - what should we do?:
ggplot(global_temps, aes(x=Year, y=TempChange)) + geom_line() + geom_point()
And adding the five-year average:
ggplot(global_temps, aes(x=Year, y=TempChange)) + geom_line() + geom_point() + geom_line(aes(y=TempChange5Year, color="red"))
But this is not nice - better to format your data to support vis - i.e., a long format:
long_global_temps = melt(global_temps, id.vars = "Year", variable.name = "MeasurementType", value.name = "Degrees")Now we can go a better job:
ggplot(long_global_temps, aes(x=Year, y=Degrees, color=MeasurementType)) + geom_line() + geom_point(data=subset(long_global_temps, MeasurementType=="TempChange"))
Histograms
Histograms are used to convey an understanding about the distribution of your data. For example, the distribution on price:
ggplot(diamonds, aes(x=price)) + geom_histogram()## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
ggplot(diamonds, aes(x=price, fill=cut)) + geom_histogram()## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
Boxplots
Box plots provide a way of showing distributions and averages. Let’s return to the airquality dataset:
ggplot(long_airquality, aes(x=Month, y=value)) + geom_jitter() + facet_wrap(~Measurement, scales="free") # the old way## Warning: Removed 44 rows containing missing values (geom_point).
ggplot(long_airquality, aes(x=factor(Month), y=value, fill=Measurement)) + geom_boxplot(varwidth = T) + facet_wrap(~Measurement, scales="free") # the old way## Warning: Removed 44 rows containing non-finite values (stat_boxplot).
What us going on with Ozone measurements for June:
subset(long_airquality, Measurement=="Ozone" & Month==6)## Month Day Measurement value
## 32 6 1 Ozone NA
## 33 6 2 Ozone NA
## 34 6 3 Ozone NA
## 35 6 4 Ozone NA
## 36 6 5 Ozone NA
## 37 6 6 Ozone NA
## 38 6 7 Ozone 29
## 39 6 8 Ozone NA
## 40 6 9 Ozone 71
## 41 6 10 Ozone 39
## 42 6 11 Ozone NA
## 43 6 12 Ozone NA
## 44 6 13 Ozone 23
## 45 6 14 Ozone NA
## 46 6 15 Ozone NA
## 47 6 16 Ozone 21
## 48 6 17 Ozone 37
## 49 6 18 Ozone 20
## 50 6 19 Ozone 12
## 51 6 20 Ozone 13
## 52 6 21 Ozone NA
## 53 6 22 Ozone NA
## 54 6 23 Ozone NA
## 55 6 24 Ozone NA
## 56 6 25 Ozone NA
## 57 6 26 Ozone NA
## 58 6 27 Ozone NA
## 59 6 28 Ozone NA
## 60 6 29 Ozone NA
## 61 6 30 Ozone NAHeatmap
A heat map is a two-dimensional representation of data in which values are represented by colors. A simple heat map provides an immediate visual summary of information. More elaborate heat maps allow the viewer to understand complex data sets.
Let’s use the diamond dataset to look at how clarity, color and price are related.
First we will do some aggregation on the data:
agg_diamonds <- aggregate(price ~ clarity + color, diamonds, mean)ggplot(agg_diamonds, aes(x=clarity, y=color, fill=price)) + geom_tile(colour = "white") + scale_fill_gradient(low = "white", high = "steelblue")